In April 2017, we posted about Dr. Mahyar Etminan, then an expert in the Mirena MDL pending in the Southern District of New York. Plaintiffs in the MDL claimed that the defendant’s product, an intrauterine contraceptive device containing the synthetic hormone levonorgestrel (“LNG”) caused them to develop idiopathic intracranial hypertension (“IIH”), also known as pseudotumor cerebri, a rare and potentially serious condition marked by increased cerebrospinal fluid pressure in the skull. In 2015, Etminan had published a study designed to assess the risk of IIH. Although the study did not definitively conclude that defendant’s product caused IIH, Etminan concluded that one of the two analyses, a “disproportionality analysis” of adverse events in the FDA’s FAERS database, identified an increased risk of IIH associated with LNG and that this result was statistically significant. Etminan concluded that the results of the second analysis, a retrospective cohort study, did not find an increased risk but that this result was not statistically significant. No other study has ever established a causal link between LNG and IIH.
Subsequently, a prominent scientist in the field attacked the methodology of Etminan’s disproportionality analysis because the study failed to control for age and gender, resulting in erroneous and misleading conclusions. At the same time, it was revealed that Dr. Etminan was on the plaintiffs’ payroll at the time that he published his study, a conflict of interest he had not disclosed. Ultimately, after defendants served Dr. Etminan with a notice of deposition in one of the cases in the MDL, Dr. Etminan repudiated much of his study’s analysis and withdrew as an expert. When we reported this, we told you to “stay tuned,” commenting that plaintiffs’ other experts, all of whom relied on Etminan’s results, had not withdrawn.
The other shoe dropped a couple of weeks ago. In In re Mirena IUS Levonorgestrel-Related Prods. Liab. Litig., 2018 WL 5276431 (S.D.N.Y. Oct. 24, 2018), the court considered the defendants’ Daubert motions to exclude the plaintiffs’ seven remaining general causation experts. And it granted them all. The opinion is very long – seventy-two pages on Westlaw – and we commend it to your weekend reading, as we can’t begin to do justice to the court’s detailed analysis of each expert’s methodology. But we wanted to bring this terrific decision to your attention and to focus its most important takeaways.
The court began its analysis by emphasizing, “In the face of [the] historical record, with no medical organization or regulator or peer-reviewed scientific literature having found that Mirena or any contraceptive product using LNG is a cause of IIH, an expert witness who would so opine . . . necessarily would break new ground in this litigation.” Mirena, 2018 WL 5276431 at *20. All seven of the plaintiffs’ general causation experts “so opined.” Four of these experts “arrived at this result largely by drawing upon existing sources.” These included varying combinations of case reports regarding Mirena, case reports regarding other contraceptive products containing LNG, another product’s warning label, the repudiated portions of the Etminan study, and another study (the “Valenzuela study”) that reported a statistically significant association between LNG-containing devices and IIH but which, the authors emphasized, found only a correlation, not a causal link. The remaining three experts were “mechanism” experts, each of whom postulated a supposed mechanism by which the defendant’s product could cause IIH. In this post, we will focus on two of the experts in the first group, which included an epidemiologist, a toxicologist, an OB/GYN, and an ophthalmologist, but we urge you to read the court’s dissection of the second group as well.
The plaintiffs’ epidemiology expert was a professor of biostatistics with experience in conducting and analyzing large clinical trials. He claimed that the nine Bradford-Hill criteria supported his causation conclusion. As many of you know, the criteria are “metrics that epidemiologists use to distinguish a causal connection from a mere association.” Id. at *23 (citation omitted). They are: statistical association (also known as “strength of association), temporality, biological plausibility, coherence, dose-response effect, consistency, analogy, experimental evidence, and specificity.
The court first held that the epidemiologist’s opinion did not satisfy any of Daubert’s four reliability factors, because the expert “has not tested his theory. He has not subjected it to peer review or had it published. He has not identified an error rate for his application of the nine Bradford Hill factors. . . . And [his theory] has not been generally accepted by the scientific community.” Id. at *27 (internal punctuation and citation omitted). With respect to this last, the court again emphasized, “Outside of this litigation, there is a complete absence of scholarship opining that Mirena, or, for that matter, any LNG-based contraceptive, is a cause of IIH.” Id. As such, the court undertook to “take a hard look” at the expert’s methodology, scrutiny that was “particularly warranted” because:
[I]t is imperative that experts who apply multi-criteria methodologies such as Bradford Hill . . . rigorously explain how they have weighted the criteria. Otherwise, such methodologies are virtually standardless and their applications to a particular problem can prove unacceptably manipulable. Rather than advancing the search for truth, these flexible methodologies may serve as vehicles to support a desired conclusion.
Id. (citations omitted). Citing four examples of how the expert’s assessment of individual Bradford Hill factors “depart[ed] repeatedly from reliable methodology,” the court held, “Measured against these standards, [the epidemiologist’s] report falls short. Id. at *28-29.
First, the expert used the “analogy” factor, basing his causation conclusion in part on an analogy to another contraceptive product. But, the court explained, this analogy was based on an “unestablished hypothesis” about the other contraceptive product, for which a causal relationship with IIH had never been substantiated. Id. at *29. With regard to the “specificity” factor, the court explained that the factor “inquires into the number of causes of a disease,” id., with the difficulty of demonstrating a causal association escalating along with the number of possible alternative causes. “In finding the specificity factor satisfied,” the expert “devote[d] two sentences to his discussion.” Id. He relied on a conclusory statement to the effect that alternative causes could be ruled out. And he relied on the Valenzuela study, which had actually disclaimed a finding of causation. The court explained that the “consistency” factor required “similar findings generated by several epidemiological studies involving various investigators” reaching the same conclusion. Id. at *30. Again, the epidemiologist claimed that the Valenzuela study satisfied this criterion because it considered two separate populations. But, as the court stated, both studies were conducted by the same investigators, and neither found a causal relationship. Finally, as to the biological plausibility factor, the epidemiologist postulated a biological mechanism by which he said LNG could cause IIH. The court stated, “ . . . [B]y any measure, [the expert] is unqualified to give an expert opinion as to a biological mechanism of causation of IIH.” Id. at *30. This lack of qualifications compromised the expert’s assessment of the biological plausibility factor as well as of related factors. The court concluded,
Each of [the expert’s] departures from settled and rigorous methodology favors the same outcome. Each enables him to find that the Bradford Hill factor at issue support concluding that Mirena is a cause of IIH. . . . [His] unidirectional misapplication of a series of Bradford Hill factors is concerning – it is a red flag. Rather than suggesting a scholar’s considered neutral engagement with the general causation question at hand, it suggests motivated, result-driven, reasoning. . . . Methodology aimed at achieving one result is unreliable.
Id. (internal punctuation and citation omitted. The court went on to further eviscerate the epidemiologist’s methodology, criticizing his reliance on the Valenzuela study, his nearly-exclusive use of case reports to support three of nine Bradford Hill factors, his failure to consider evidence that undercut his opinions, and his cherry-picking of case reports that supported his desired conclusion. The court concluded that the expert’s testimony was “compromised by a range of serious methodological flaws,” and failed to satisfy Daubert’s reliability standard.
The court voiced similar criticisms of the methodology of the plaintiffs’ toxicology expert. Like the epidemiologist, the toxicologist failed to meet any of the four Daubert reliability standards In applying the Bradford Hill factors, she failed to identify support for her conclusions, distorted or disregarded evidence that undercut her opinions, failed to articulate a plausible biological mechanism to support her causation conclusion, and drew an inapposite analogy to another contraceptive product. And her opinions were plagued by additional methodological flaws. She relied on the portion of the Etminan study that was discredited and that Etminan himself repudiated. And she cited the Valenzuela study as her sole support for finding several Bradford Hill criteria satisfied without acknowledging the study’s methodological limitations and failure to find causation. The court concluded, “[The toxicologist’s] proposed testimony is beset by methodological deficiencies. It falls far short of satisfying Daubert’s standard of reliability. Her testimony, too, must be excluded.” Id. at *40.
And so it went with the court’s discussion of the rest of the plaintiffs’ experts. The opinion does the best job we’ve ever seen of demonstrating how an expert can attempt to create the illusion of reliability by paying lip service to the Bradford Hill criteria and how those criteria can be manipulated to mask wholly result-driven ipse dixit opinions plagued by fatal methodological flaws. In this case, a committed and rigorous judge stemmed the tide. But we all know that this is not always the case.
We love this decision. There is a lot more to say about it, and we look forward to telling you more in an upcoming post.